[pgpool-general: 2436] Re: pgpool connections: FATAL: sorry, too many clients already

[Sergey Arlashin]

On Jan 16, 2014, at 10:42 AM, Tatsuo Ishii <ishii at postgresql.org> wrote:

>> On Jan 16, 2014, at 10:10 AM, Tatsuo Ishii <ishii at postgresql.org> wrote:
>> 
>>>> It seems I misunderstood this formula :) 
>>>> 
>>>> I have 2 backends. Each can handle 27 connections. So I want pgpool to be able to handle 27*2 connections. 
>>>> 
>>>> 
>>>> If I use 
>>>> 
>>>> 13 * 2 <= (30 -3 ) 
>>>> 
>>>> then the total number of connections won't exceed 26. Is that correct? 
>>> 
>>> Yes, correct.
>>> 
>> 
>> So this where this calculation comes from. 
>> I mean: 25*2 <= (30-3)*2
> 
> No. Pgpool-II connects to each backend independently. That's the
> reason why following formulas do not include the number of PostgreSQL
> nodes.
> 
> max_pool*num_init_children <= (max_connections - superuser_reserved_connections) (no query canceling needed)
> max_pool*num_init_children*2 <= (max_connections - superuser_reserved_connections) (query canceling needed)
> 
>> If I use 13*2 <= (30-3) then there will be no way to utilize the capacity of all backend servers. Is that so? 
> 
> That depends on the definition of "utilize". Pgpool-II utilizes
> standby servers as a resouce for load balance. It does not utilize
> max_connection of PostgreSQL.
> 

So in other words even if I have 10 backends with max_connections = 30 each, I still should use 
num_init_children = 13
max_pool          = 2
in pgpool.conf. 

Is that correct? 








> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_en.php
> Japanese: http://www.sraoss.co.jp